Forest Ecology and Management 419–420 (2018) 1–9

Contents lists available at ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier.com/locate/foreco

An innovative approach to disentangling the effect of management and T environment on cover and density of protected areas in African

Mario Messinaa, Robert Cunliffea, Alessio Farcomenib, Luca Malatestaa, Izak P.J. Smitc, ⁎ Riccardo Testolina, Natasha S. Ribeirod, Bruno Nhancalee, Marcello Vitalea, Fabio Attorrea, a Department of Environmental Biology, Sapienza University of Rome, P.le A. Moro 5, Rome, Italy b Department of Public Health and Infectious Diseases, Sapienza University of Rome, Italy P.le A. Moro 5, Rome, Italy c Scientific Services, South African National Parks, , Private Bag X402, Skukuza 1350, d Departamento de Engenharia Florestal, Universidade Eduardo Mondlane, P.O. Box 257, Maputo, e Departamento de Biologia, Universidade Eduardo Mondlane, P.O. Box 257, Maputo, Mozambique

ARTICLE INFO ABSTRACT

Keywords: In protected areas of the African savanna tree cover, structure and species composition are influenced by a Collect Earth combination of many different variables. These include complex and multi-scaled interplay of environmental Fire frequency factors such as water and nutrient availability, fire, herbivory and, when occurring, direct human disturbance. In Kruger National Park this study, we conducted a comprehensive and comparative analysis of the spatial variability of tree cover and Gonarezhou National Park density in three neighboring Southern African National Parks (Kruger, , and Gonarezhou) characterized Limpopo National Park by similar environmental conditions but different management plans. We sampled 3382 plots of 0.5 ha across the Soil three parks using an innovative methodology defined as augmented visual interpretation, based on a free and open source software. This software, named Collect Earth, allows access to very high spatial and temporal resolution imagery archives. Spatial variability of tree cover and density was analyzed comparing the three parks and the two bioclimatic regions (semiarid and dry subhumid) characterizing them. The effect of relevant en- vironmental variables such as edaphic factors, precipitation and fire frequency was also assessed. Kruger National Park is characterized by the lowest values of tree cover and density among the three Parks. Contrary to what was expected and the general trend of Southern Africa, the dry subhumid zone showed lower values of tree cover and density than the semiarid zone. Such variability is hypothesized to be related to the different man- agements of the three parks within the general environmental template characterizing the African savanna as well as differences in tree species composition between the two climatic zones.

1. Introduction tree seedlings, which in turn have negative effects on the establishment of the latter (Kraaij and Ward, 2006). Within this environmental tem- In protected areas of the African savanna the cover, structure and plate, the synergistic effect of fire regime and density of mega-herbi- composition of woody vegetation are affected by the complex and vores (mainly , Loxodonta africana) can have a significant multi-scale interplay of environmental factors such as water and nu- impact on woody cover and structural diversity (Van Langevelde et al., trient availability, fire, herbivory and, when occurring, direct fuel wood 2003; Holdo et al., 2009; Asner and Levick, 2012; Levick and Asner, harvesting by people (Van Langevelde et al., 2003; Holdo, 2007; 2013). Natural and prescribed fires have a prominent role in main- Sankaran et al., 2008; Shannon et al., 2011; Vanak et al., 2012; taining the equilibrium between grasslands and woodlands in savanna Buitenwerf et al., 2012; Holdo et al., 2013). At regional and landscape ecosystems by reducing woody vegetation cover and density (Govender scales, water availability, mainly related to rainfall regime, is con- et al., 2006), especially in areas with higher precipitation rates (Devine sidered the primary resource driver (Kerkhoff et al., 2004; Sankaran et al., 2015). Similarly, the presence of herbivores generally leads to a et al., 2005)defining the maximum potential tree cover (Coughenour reduction in woody cover due to browsing, grazing (Staver et al., 2009; and Ellis, 1993; Sankaran et al., 2005). As for soil nutrients (i.e. ni- Staver et al., 2011) and physical damage (Asner and Levick, 2012). trogen), they are more limiting to grasses than to : a high soil ni- Despite the evolution of traits to resist or tolerate these dis- trogen content promotes the growth of herbaceous species rather than turbances, such as re-sprouting ability, storage in below ground organs,

⁎ Corresponding author. E-mail address: [email protected] (F. Attorre). https://doi.org/10.1016/j.foreco.2018.03.019 Received 16 November 2017; Received in revised form 12 March 2018; Accepted 13 March 2018 0378-1127/ © 2018 Elsevier B.V. All rights reserved. M. Messina et al. Forest Ecology and Management 419–420 (2018) 1–9 and fast height and diameter growth rates (Higgins et al., 2000; Julbernardia globiflora. The mopane vegetation type is dominated by Midgley et al., 2010), fire (especially high intensity fires – Smit et al., Colophospermum mopane and apiculatum (Martini et al., 2016) and elephants are a lethal combination for woody limiting 2016). recruitment and transition processes (Levick et al., 2009; Midgley et al., 2010; Shannon et al., 2011). This effect is particularly evident in certain 2.1.2. Limpopo National Park protected areas where the increasing density of populations is Limpopo NP (Mozambique) was proclaimed in 2002 and covers an causing a significant decline of large trees (Asner and Levick, 2012; area of 1 000 000 ha. The Kruger National Park in South Africa neigh- Levick and Asner, 2013; O’Connor and Page, 2014). Given the range of bors the Limpopo NP to the west, the Limpopo River forms the northern ecosystem services that woody vegetation provides (e.g., soil protec- and eastern border, whereas the Olifants River (called Rio dos Elefantes tion, food for large herbivores, carbon sequestration) and the cascade in Portuguese) forms the southern boundary. Elevation ranges from effect that its change exerts on several important features of savanna 521 m asl in the north down to 45 m asl at the confluence of the ecosystems such as nutrient patterns (Ludwig et al., 2004; Treydte Limpopo and Olifants rivers. Rainfall decreases from 500 mm near the et al., 2007), habitat suitability (Cumming et al., 1997; Parr and Massingir Dam in the south to < 450 mm at Pafuri in the north. The Andersen, 2006), herbaceous biomass and fire (Smit and Prins, 2015), dominant geological feature of the Limpopo NP is the extensive sandy parks managers need to be supported by monitoring tools in order to cover along the northwest/southeast spine of the park. Calcareous se- allow for rapid and comprehensive woody vegetation assessments. In dimentary rocks have been exposed where this sand mantle has been this context, remote sensing has been recognized as a fundamental tool eroded closer to the main drainage lines. Alluvial deposits are found for ecologists and a primary source of data for managers and policy along the main drainage lines (Limpopo, Olifants and Shingwedzi). makers. However, the application of conventional analysis techniques Vegetation is characterized by Mopane woodlands dominated by (e.g. semi-automated image classification) has proven to be challen- Colophospermum mopane, Combretum apiculatum and Terminalia sericea, ging, especially in combination with very high resolution (VHR) sa- and other more localized types such as Acacia tortilis and Acacia xan- tellite imagery, due to its high costs, low spatial extent and lack of thophloea forests along the Limpopo River, Terminalia prunioides thickets global coverage (Bey et al., 2016). In this study, we used an innovative on shallow, stony soils and fragmented stands with Androstachys john- methodology (augmented visual interpretation) based on a free and sonii on steep calcrete slopes (Stalmans et al., 2004). open source software, Collect Earth, developed by the Food and Agri- culture Organization of the United Nations (FAO). Such technology 2.1.3. Kruger National Park allows to perform land cover assessments through visual interpretation Kruger NP (South Africa) covers almost 2 000 000 ha. The altitude of freely accessible VHR satellite imagery archives (Bey et al., 2016). ranges from 200 m in the east to 840 m asl in the south-west. There is a The analysis was carried out in three national parks (NPs), which marked increase in mean annual precipitation from the north to the are part of the Great Limpopo Transfrontier Conservation Area: Kruger south of the park ranging from about 440 mm in the north to about National Park (South Africa), Limpopo National Park (Mozambique) 740 mm in the south (Venter and Gertenbach, 1986). The park is geo- and Gonarezhou National Park (), characterized by similar logically divided into two main parts; granites and their erosion pro- environmental conditions but different management capacities. Indeed, ducts to the west and basalts and their erosion products to the east. The while Kruger NP is characterized by highly developed infrastructures vegetation in Kruger NP can be divided into 35 vegetation types, with and established wildlife and fire management regimes, Limpopo NP has the woody component largely dominated by Colophospermum mopane none and Gonarezhou NP represents an intermediate situation. The (in the north of the park), Combretum apiculatum, Acacia nigrescens, aims of the study are: (1) to assess the spatial distribution of woody Acacia tortilis, Terminalia sericea and (Gertenbach, vegetation cover and density in the three parks by applying a novel 1983). methodology and (2) to estimate the relative influence of environ- mental drivers (i.e. precipitation and edaphic variables) and dis- 2.2. Data set turbance factors (i.e. fire frequency) on woody vegetation in the context of different management capacities. Data collected represent a baseline In February 2016, 3382 plots of 0.5 ha were sampled using the that can be used to assess future changes and the outcomes of the im- augmented visual interpretation approach introduced above and de- plementation of management strategies focused on fire, wildlife, cli- scribed in detail in the following paragraphs. The plots were homo- mate change, and in the case of the Limpopo NP and Gonarezhou NP, geneously distributed on a 3 km square-cell regular grid laid over the respectively on old and newly established settlements. entire area of each NP, thus the number of plots sampled for each NP is proportional to its area (Kruger = 1875, Limpopo = 1031, 2. Material and methods Gonarezhou = 476). The plot size of 0.5 ha was chosen to be consistent with the FAO-FRA definition of forest, which has a tree cover ≥10% 2.1. Study area spanning an area of more than 0.5 ha that is not predominantly used for agriculture or urban activity, as well as areas in which tree cover is 2.1.1. Gonarezhou National Park temporarily < 10% but is expected to recover (FAO, 2001). Gonarezhou NP (Zimbabwe) was established in 1975 in the south- The assessment was conducted using Open Foris Collect Earth, a east of the country alongside the border with Mozambique, stretching new tool developed by FAO and based on recent developments in cloud between the Mwenezi and Save Rivers. It comprises a roughly rectan- computing systems, such as Google Earth Engine and the increasing gular strip of land of about 35 to 45 km wide and 135 km long, with an availability of VHR satellite imagery. This innovative application was area of about 500 000 ha. Altitude ranges from 160 m asl at the Save- used for a global assessment of dryland forests (Bastin et al., 2017) and Runde junction to a maximum of 578 m asl on the Chiwonja hills. Mean allows the assessment of land cover through visual interpretation of annual rainfall over a 29-year period (from 1961 to 1990) was 515 mm. remote sensing imagery from DigitalGlobe libraries, accessed through Morphologically, the Gonarezhou NP forms part of the Limpopo-Save Google Earth and Microsoft Bing Maps. The present study is based on Lowlands of Zimbabwe, extending across the southernmost part of the interpretation of satellite imagery available at the time of the assess- country in the form of a relatively flat plain that rises gently to the ment (between February and March 2016) with a spatial resolution north from the Limpopo River. Vegetation is characterized by Miombo from ≤1 m (VHR; ∼96% of plots) to ≤10 m (high resolution, e.g. and Mopane vegetation types. The former is dominated by Brachystegia SPOT, RapidEye and Sentinel 2; ∼4% of plots) and to > 0–100 m tamarindoides subsp. torrei, Combretum celastroides subsp. celastroides, (medium resolution, e.g. Landsat; < 1% of plots), coupled with the Combretum collinum subsp. collinum, Guibourtia conjugata and analysis of the Normalized Difference Vegetation Index (NDVI)

2 Download English Version: https://daneshyari.com/en/article/6541699

Download Persian Version:

https://daneshyari.com/article/6541699

Daneshyari.com